Social media firm Snap said on Monday it is rolling out an experimental chatbot feature running on OpenAI’s GPT technology for some users of its photo messaging app Snapchat. The feature dubbed as My AI will be running the latest version of OpenAI’s GPT technology which has been customised for Snapchat. The AI chatbot will be initially rolling out to Snapchat+ subscribers as an experiment, releasing this week. The social media platform has also requested users to submit their feedbacks to improve the chatbot for future application. 

In an official blog post, Snap announced the launch of its experimental AI chatbot that works on the latest version of OpenAI‘s GPT technology. The chatbot or My AI will be rolling out this week and will be limited to the subscribed users on Snapchat+ for now. However, this may roll out to all Snapchat users in the coming months. The company has not hinted at any such development in its blog. 

Snap has clarified that it is possible for My AI, in its initial phase, to commit mistakes, but the company aims to avoid any “biased, incorrect, harmful, or misleading information” through the chatbot. Similar to other chatbots like Microsoft‘s ChatGPT, or Google Bard, My AI can be tricked into revealing unwanted information. 

However, the company will be saving all conversations to review the experimental chatbot. Based on the review and feedback from the users, the company will introduce further changes for improvement. Snap has also requested its users to avoid sharing secrets and personal information with the AI chatbot. 

Last week, Meta announced a research tool for building artificial intelligence-based chatbots and other products, named as LLaMA. The tools is being planned to be made available to AI researchers soon. However, it is currently not available on Meta-owned apps such as WhatsApp and Facebook.